Note: Descriptions are shown in the official language in which they were submitted.
CA 02297658 2000-O1-21
WO 99/05584 PCTlSE98I01316
1
INTERNET CACHING SYSTEM
Technical field of invention
The present invention relates to a method, a system
and a server for caching Internet information content.
Background to the invention and Prior Art
Over the last few years, Internet has developed into
the fastest growing means of communication and is predic-
ted to be the main system for distributing information in
the future. The Internet has become increasingly popular
and the number of users is growing at a magnificent rate.
Anybody can provide any information - text, pictures,
audio and video - on the net where it can be retrieved by
users anywhere in the world. This makes for the incredi-
ble success of the Internet and its currently most used
feature, the World Wide Web, WWW.
However, the popularity of the Internet, and
especially the use of the Internet application World Wide
Web, is putting the net under enormous capacity pressure.
Traffic on the Internet doubles about every three months,
with no signs of slowing down. This kind of exponential
growth will eventually slow down, but several factors
indicate that it is still far from any slowdown in
growth.
Today, only a small fraction of all PCs have access
to the Internet. It is believed that most of those who
have made and will make the investment in a PC will
eventually want access to the Internet. Furthermore,
communication speeds between the user and the Internet
infrastructure is increasing rapidly. Technologies such
as ISDN, Cable TV modems and xDSL are emerging. Moreover,
manufactures of household electronics are launching low
costs devices to give Internet World Wide Web access
through television sets.
CA 02297658 2000-O1-21
WO 99/05584 ~ PCT/SE98/01316
2
These and other factors are all contributing in
making it increasingly difficult to access information on
the Internet, since the system is basically being over-
loaded. In many parts of the world, the lack of suffi-
cient bandwidth is a major problem. Also, the passing
information between USA and Europe is limited by the
transmission capacity of the transatlantic link. The
transmission capacity is simply too small to cope with
the exploding Internet use.
Basically, there exists two conventional solutions
to this problem. A first solution is to add more band-
width and switching capacity, which is being done, but
which carries huge costs and also faces significant
technical problems.
A second solution is to use so called caching tech-
niques. Caching basically means to keep track of the
Internet traffic and to keep copies of the most frequent-
ly accessed World Wide Web files at a location closer to
the user than the original files. This hence involve the
storing of a local copy of a World Wide Web Site of
interest. For example, a copy of CNN's USA-based homepage
may be temporarily stored in a cache at a European
"proxy" location, whereby European Internet users may
access the CNN homepage without having to use the trans-
atlantic communication link, thereby both attaining
quicker access and at the same time lowering the load on
the net. The total World Wide Web information content
itself is getting to be unmeasurable, but only a rela-
tively small subset of all this information accounts for
a huge proportion of what is actually viewed. Depending
on the size and homogeneity of a user community, about
10-20 Gigabytes of cache storage will (spring 1997)
reduce the traffic by 30-50a in that community.
However, in spite of current caching techniques, in
many parts of the world, such as South and East Europe,
South America, India and East Asia, the bandwidth
CA 02297658 2000-O1-21
WO 99105584 PCTJSE98/01316
3
capacity of the available communication systems are far
too low to provide a desired usability of the Internet.
Summary of the invention
The object of the invention to reduce the capacity
constraints on the Internet by providing a caching solu-
tion which reduces access time on the Internet and which
allows growth to a significantly larger amount of users
and information traffic on the Internet, especially in
areas where the bandwidth capacity are low.
According to the present invention, said object is
achieved by a method, a system and a server for caching
Internet information content according to the appended
claims.
The invention is based upon the insight that a
linguistically and culturally homogenous geographical
area often differ in size and shape from the economically
and technically optimal region to be served by a single
cache server. Therefore, such an area is served by a set
of geographically distributed servers. However, the
invention is further based upon the insight that, under
the assumption that users in such a linguistically and
culturally homogenous geographical area will have
homogenous preferences regarding visited information
sites, information cached in one such server due to a
user request should also be cached in the other servers
within said area, the basic principle being that if one
user is interested in a certain piece of information, it
is likely that so will other persons within the same
linguistically and culturally homogenous geographical
area. The larger the end user community, the larger the
probability that someone else within the community has
requested any given file.
Hence, according to the invention, a set. of geo-
graphically distributed cache servers serving different
geographical regions but having a common relation, such
as the relation of serving a culturally and/or linguisti-
CA 02297658 2000-O1-21
WO 99/05584 PCT/SE98/01316
4
cally defined user group or area, is updated with essen-
tially the same information whenever one of said cache
servers retrieves information due to an information
request from an end user or the like.
Such cultural and linguistical homogeneity may be
defined by the boundaries of a nation, the boundaries of
a specific language, the boundaries of a specific
religion, the boundaries of a certain level of technology
or development, the boundaries of a certain economical
area, as well as combinations thereof, or the like.
Hence, according to an embodiment of the invention,
said set of geographically distributed cache servers are
distributed within a linguistically and culturally
defined global geographical area or a geographically
defined community. Furthermore, each of said cache
servers is preferably arranged to serve a respective
local region within said geographical area or community.
For example, the German speaking parts of Germany,
Austria, Switzerland and Italy may be seen as defining a
cultural and linguistically homogenous global area.
Therefore a set of cache servers operating according to
the invention is distributed within this area. Among this
set of servers, provided within the global area defined
above, one (or more) may is for example arranged to serve
a first region covering the northern part of Germany, a
second one (or more) to serve a second region covering
the eastern part of Germany, a third one (or more) to
serve a third region covering the central part of
Germany, a fourth one (or more) to serve a fourth region
covering Switzerland and the northern part of Italy, and
finally a fifth one (or more) to serve a fifth region
covering Austria.
Note that a defined geographical area according to
the invention does not necessarily imply a homogenous,
continuos area encompassed by a closed boundary, but may
just as well mean a set of geographically separated sub-
CA 02297658 2000-O1-21
WO 99/05584 PCT/SE98I01316
areas together defining an area according to the inven-
tion.
According to estimations made by the inventor, in
some parts of the world, such as India and South America,
5 the bandwidth capacity of the telecommunication network
is so small, that it is actually more efficient to
distribute a specific package of information to all cache
servers within the defined global area based upon just
one information request to the net from one single user,
than to have another user search the net for the same
package of information a second time. However, rules for
determining when to perform distributed caching according
to the invention or not, may be set in different ways
depending on the actual application environment, as will
be discussed below.
The distribution of information to the set of cache
servers according to the invention may be performed in
many different ways. According to one aspect, the
Internet itself could be used. However this would put
event more strain on the transfer capacity of the net.
Therefore, according to a preferred embodiment of the
invention, the information is distributed to said set of
cache servers using multicast communication, preferably
using dedicated data channels separated from the Internet
communication itself.
According to a preferred embodiment, said informa-
tion is distributed to said set of cache servers using a
satellite link. Each of the servers then preferably have
access to an uplink for sending said information, and all
servers receives a downlink which provides information
sent from any one of the servers. The hardware and soft-
ware for handling such distribution may be provided in
association to the cache server itself, of may be pro-
vided by for example an Internet Service Provider or an
Satellite Communication Link Provider. Similarly, the
software for controlling the information flow and for
performing decisions within the system may be provided at
CA 02297658 2000-O1-21
WO 99/OS584 PCTISE98I01316
6
the cache servers, at the Internet Service Providers or
wherever convenient, as will be clear for a person
skilled in the art. If a satellite system is used, the
satellite link may also be used for transferring or
relaying request to other parts of the globe, for example
from Europe to USA and vice versa.
According to an alternative embodiment, all cache
servers within the system are connected via data channels
to a central control unit. Such a central control unit
may for example control the decisions as to weather or
not information should be cached and as to how it is to
be distributed and so on.
Depending on the size of the community, the capacity
of the cache servers, the available communication capa-
city~of the net, the objects of the cache provider, and
so on, the distributed caching technique according to the
invention may be arranged to distribute, or not distri-
bute, information based on different rules. According to
a preferred embodiment of the invention, only requests or
frames referring to a specific communication format or
application, corresponding to the type of information
service provided on the local storing means, is cached.
For example, only so called TCP-queries to the WWW port.
According to another example, the cache server provider
may decide only to cache from specific addresses (e. g.
only .com or .org addresses), only after a certain number
of requests has been made within the system (e. g. after
three separate requests), or the like.
The processing means necessary for performing such
decisions may be arranged at each cache server, at a
central control unit as mentioned above, at an Internet
Service Provider or similar location working in associa-
tion with the cache servers, or the like.
Also request relating to an Information Content
Provider located in a region served by one of the cache
servers in said set of cache servers may be handled in
CA 02297658 2000-O1-21
WO 99/05584 PCTlSE98/01316
7
different ways according to different embodiments of the
invention, as will be discussed in the following.
.According to two first embodiments thereof, it is
assumed that a local information request from an end user
to an Internet content provider, both being located
within the same region, only is of local concern. There-
fore the information shall: a) not be cached at all,
since a local user will always have easy access to a
local Internet Content Provider; or b) only be cached in
the local cache server and not be distributed to the
entire set of geographically distributed cache servers,
since the request probably only is of local concern.
Hence, in this context, the term "local" refers to
features residing within one of said regions.
According to a second embodiment, it is assumed that
an information request from an end user to wn Internet
content provider located within one of said regions shall
not be cached in the cache server serving that region,
but shall instead only be distributed to all other cache
servers within the system, since a local user will always
have easy access to a local Internet Content Provider
(hence no need for local caching) but a user outside the
region will not have the same easy access to the local
Internet Content Provider (hence the need for caching
outside the region).
According to another embodiment, it is assumed that
the cache server capacity is large enough to handle all
request, therefore there is no need for handling informa-
tion relating to local Internet Content Providers diffe-
rent than other information and, consequently, the
caching of information in one cache server within the
system shall result in the caching of the same informa-
tion in all cache servers within the system.
According to yet another embodiment of the inven-
tion, depending on the choice of rules for decisions
within the system, there is provided a list of Internet
Content Provider addresses to be excluded from caching.
CA 02297658 2000-O1-21
WO 99105584 PCT/SE98101316
8
Hence, each cache is arranged to check incoming informa-
tion requests against said list to see whether or not the
request shall lead to caching.
Depending on the desired utility, the cache servers
will be arranged to store information for a given amount
of time. For example, the caches may operate on the
principle of first-in-first-out. According to another
example, certain kind of information may be provided with
update requirements, which will define the longest time
interval that the copied information may be used to
ensure that the information is still correct. Of course,
regarding this and other aspects, the invention may be
combined with different types of conventional caching
techniques, as is understood by one skilled in the art.
Furthermore, each of said cache servers may be
constituted by several sub-servers connected, preferably
via high speed switching means, to a cache manager
controlling the operation thereof. Further, there is no
limit to the number of cache servers that may be used.
Also, each server may be used by more than one Internet
Service Provider.
In the event of a failure of one of said cache
servers, it can rapidly be bypassed by the system to
restore some service while the cache server is being
repaired or replaced. The cache functionality may then be
provided form another server in another region within the
system. An advantage of the invention is that the auxili-
ary cache server then provides the same stored content as
the bypassed serves, thus making sure that the cached
information is still relevant to the end users in the
region of the bypassed server.
Although the description of the invention has been
made with respect to the Internet communication system,
it is understood by those skilled in the art that the
invention also may be used in other similar types of
global information communication systems which may be
found in the future and which may show similar kinds of
CA 02297658 2000-O1-21
PCT/SE98I01316
9
communication problems. The invention is therefore of
course not limited to the Internet application.
Also, even though the description further has been
made mainly with respect to the World Wide Web applica
tion, it is to be understood that the invention may as
well be used for many other kinds of net applications and
is not limited thereto.
_Brief description of the drawings
Further aspects, features and advantages of the
invention will become clear from the following descrip-
tion of preferred exemplifying embodiments with reference
to the accompanying drawings, in which:
Fig 1 schematically shows an embodiment of an inter-
net caching system according to the present invention:
Fig 2 schematically shows a cache server arrangement
forming a part of the system shown in Fig 1;
Fig 3 schematically shows an alternative cache
server arrangement
Fig 4 schematically shows yet another alternative
cache server arrangement
Fig 5 schematically shows a flow chart of the
operation performed by the cache server in Fig 4;
Fig 6 schematically shows a flow chart of the
operation performed by the interceptor in Fig 3;
Fig 7 schematically shows another embodiment of an
Internet caching system according to the present
invention; and
Fig 8 schematically shows yet another embodiment of
an Internet caching system according to the present
invention.
_D_etailed description of preferred embodiments
A first embodiment of an Internet caching system
according to the invention will now be described with
reference to Fig 1. In Fig 1, a preferably linguistically
and/or culturally defined geographical area 200 is shown
CA 02297658 2000-O1-21
WO 99/05584 PCT/SE98/01316
encompassing four geographical regions A, B, C and D. For
example, the area 200 could be the British islands, and
the regions A, B, C and D could then be Ireland: Wales:
Scotland and Northern England: and Southern England,
5 respectively.
In each of the regions A, B, C, and D, there is at
least one Internet Service Provider 110a, 110b, 110c, and
110d, respectively, providing the respective region with
access to the Internet 100. Each service provider 110a,
10 110b, 110c, 110d, hence provides Internet connection to a
number of end users 120a, 120b, 120c, and 120d, respec-
tively, and a number of Internet content providers 130a,
130b; 130c, and 130d, respectively. It is understood
that, even though the end users 120 and content providers
130 are shown in Fig 1 as being separate features, an end
user may be a content provider as well.
Each service provider is connected to at least one
Internet cache server 140a, 140b, 190c, 140d, respec-
tively. Each cache server 140 hence provides cache
functionality to the end users 120 within the respective
region. Also, each cache server 140a, 140b, 140c, 140d,
has a connection to a satellite disc 150a, 150b, 150c,
and 150d, respectively, and may broadcast information to
all other cache servers within the area 200 using a
satellite 160.
An example of the operation of the system shown in
Fig 1 will now be described. When, for example, a service
provider 110a in region A receives an information request
from an end user 120 relating to information provided by
a content provider 130e located in a region E, which in
this case lies outside the area 200 (for example, the
region E resemble North America), the service provider
110a checks its associated cache server 140a to see if an
updated copy if the requested information is stored
therein. If so, the requested information is derived from
the cache server 140a and is then delivered to the end
user. If, however, no copy of the requested information
CA 02297658 2000-O1-21
WO 99105584 PCTISE98101316
11
is stored in the cache server 190a, the service provider
110a~relays the request to the original content provider
130e via the Internet 100. When the requested information
is returned from the content provider 130e via the
Internet 100, the service provider 110a delivers the
information to the end user and updates the cache server
140a with the received information. As a part of the
updating operation, the updated information is sent via
the satellite disc 150e and the satellite 160 to the
cache servers 140b, 140c, 140d in the other regions B, C,
and D, thereby updating all servers with the same
information. After this updating has been made, no user
120 within the area 200 need to access the original
content provider 130e via the Internet, but can instead
use the copy stored in the respective cache server. If
any cache server is already full when updating, the
information having been stored the longest period of time
(since last accessed by an end user) will be deleted from
said server. Also each server will use on of several
methods to ensure that the information stored in each
cache server are accurate and up to date, hence providing
reliable copies of the original sites.
In the following figures, for ease of description,
elements having similar functions as those described with
reference to Fig 1 will be designated with the same
numerals.
A cache server arrangement forming part of a system
of the kind shown in Fig 1 will now be described with
reference to Fig 2. In Fig 2, three end user terminals or
stations 120 are connected to a multiplexing unit 170,
which may be a modem pool or a LAN router depending on
the type of user system. The multiplexing unit 170 is
connected to an interceptor 110 arranged at an Internet
Service Provider providing access to the Internet 100,
The multiplexing unit 170 multiplexes the communications
to and from the user terminals 120.
CA 02297658 2000-O1-21
WO 99/05584 PCT/SE98/01316
12
According to this arrangement, a user 120 wishing to
access information provided for example as a World Wide
Web homepage or site by an Internet content provider (not
shown), located at a original site of publication some-
where on the Internet 100, generally sends an information
request, specifically a HTTP request to the WWW-port, via
the multiplexing unit 170 to the Internet service provi-
der 110, where the request is routed out on the Internet
100 to finally reach the content provider. The content
provider then answers the request, sending the requested
information via the Internet 100 back to the user 120 via
the service provider 110.
However, according to the arrangement shown in Fig
2, the information request from a user 120 is examined by
the interceptor 110 at the service provider. The inter-
ceptor then decides whether or not the requested informa-
tion, for example the information provided by said the
content provider, exist in a copy or similar form on a
local cache server 140 arranged in direct communication
with.the interceptor 110. If a copy or the like of the
requested information is stored at the local server 140,
the request from the user is re-routed to the server 140,
which will then return the requested information to the
user. In this case, the information will already exist on
all cache servers within the area 200 shown in Fig l, and
there is hence no need for distributing the information
to other caches.
However, if the interceptor 110 determines that no
copy or the like of the information of interest exists on
the local server 140, then the information request from
the user 120 is sent on to the Internet 100 as in the
conventional form of Internet communication. The inter-
ceptor will then monitor the response from the Internet
100 to the user 120 and update the cache server 140 with
the requested information as it is received.
The caching of new or updated information in the
cache server 140 is, according to the invention, accompa
CA 02297658 2000-O1-21
WO 99/05584 PCT/SE98/01316
13
vied by the distribution of the new or updated informa-
tion to other cache servers within the area 200 shown in
Fig 1. This distribution is enabled by either the cache
server 140 or the interceptor 110 at the service provi-
der, for example, and uses a satellite uplink to the
satellite 160 to reach other cache servers within the
defined area. Similarly, when receiving information
relating to new or updated information from other cache
servers within the area, the cache server 140 will rece-
ive such information via a downlink from the satellite.
Note that the interceptor 110 in this example only
intercepts communication packages of certain specific
types. For example, only World Wide Web information
requests from the users 120 is intercepted. Other types
of communication services, such as e-mail or the like, is
left undisturbed by the interceptor 110.
An alternative cache server arrangement, similar to
the one shown in Fig 2, will now be described with
reference to Fig 3. The arrangement in Fig 2 differs from
the one shown in Fig 2 in that the interceptor 110 at the
service provider is connected to not only the cache
server 140, but also to a quick access cache memory 112.
In comparison to the cache server 140 according to the
invention, for example providing several hundred Giga-
bytes of storage, the quick access cache memory will in
this case provide a much smaller storage capacity, for
example a around one Gigabyte of storage. The cache
memory 112 is therefore used to store essentially all
information passing the interceptor 110, but
consequently, each package of information will only be
stored a comparatively small amount of time.
In Fig 3, the quick access cache memory, comprising
a memory areas 112 and a look-up table 111 showing which
information is currently stored in said memory areas, is
used as a working memory of the interceptor 110. Hence,
when receiving an information request from a user 120,
the interceptor will first check its quick access cache
CA 02297658 2000-O1-21
WO 99105584 PCT/SE98/01316
14
memory 111, 112 for the requested information. The cache
server 140 will then be accessed only if the requested
information cannot be found on the cache memory 112. The
operation the follows the one described with reference to
Fig 2. However, in Fig 3, all information relating to
every information request from the users 120 will be
temporarily stored in the quick access cache memory 112,
either information received from the Internet 100 or
information received from the cache server 140, whereas
the cache server will only be updated when new informa-
tion is received from the Internet 100. Furthermore, the
cache memory may be user to cache information of other
kind than the information being stored in the cache
server 140. For example, in addition to storing WWW
information, as the cache server 140, the cache memory
may also store information relating to other types of
queries. Also, information stored in the cache memory 112
will not automatically be distributed to other caches
within the area 200 in Fig 1.
Another alternative cache server arrangement, also
similar to the one shown in Fig 2, will now be described
with reference to Fig 4. In Fig 4, all WWW requests will
be relayed form the service provider 110 to the cache
server 140. The cache server will then answer the
- 25 request, if the requested information is stored in there-
in, or itself handle the information retrieval from the
intexnet 100, either using a satellite link or some other
connection 141 to the Internet. When receiving the
requested information, the cache server will 140 update
itself, as well as other caches via the satellite 160,
according to its own decisions, and deliver the informa-
tion to the end user 120 via the service provider 110.
Note in Fig 4 that the service provider may still
handle other type of services, such as email or the like,
on its own, without relaying such information packages
via the cache server.
CA 02297658 2000-O1-21
WO 99/05584 PCT/SE98/01316
An example of the operations performed by the cache
server 140 in Fig 4 will now be described with reference
to the flow chart in Fig 5. After activating the opera-
tion in step S1, the cache server receives, in step S2,
5 an information request from an end user relating to a
specific content provider. In step S3, the cache server
then determines whether or not a copy of the requested
information is stored therein. If so, the requested
information is delivered to the end user in step S4 and
10 the operation is ended in step 510. (Of course, the end
step S10 may cause a return to the starting step S1).
However, if it is determined in step S3 that no copy
of the requested information is stored in the cache
server, the process continues t.o step S5, wherein the
15 cache server sends a request for said information to the
designated content provider via the Internet. The cache
server then receives the requested information in step S6
and delivers it to the end user in step S7. In step S8,
the received information is stared in the cache server,
and then, in step S9, the information is distributed to
all other cache servers within the area 200 shown in Fig
1. Hence, the received new, previously not stored
information will then be stored in essentially all cache
servers within the area and thus be easily accessed by
all users within the area.
An example of the operations performed by the inter-
ceptor or service provider 110 in Fig 3 will now be des-
cribed with reference to the flow chart in Fig 6. After
activating the operation in step S11, the interceptor 110
reads, in step S12, an information request from an end
user directed to a specific content provider. In step
513, the interceptor 110 checks the content table 111 of
the quick access cache memory 112 to see if a copy of the
requested information exists therein. If so, the opera-
tion continues to step S15 described below. In not, the
operation continues to step 514, in which the interceptor
sends a query to the cache server 140 to see if a copy of
CA 02297658 2000-O1-21
WO 99/05584 PCT/SE98/0131G
16
the requested information exists therein. If the answer
is yes, the process continues to step 515, in which the
request from the end user is intercepted, and to step
516, wherein the requested information is delivered,
either using the cache memory 112 or the cache server
140. However, if the answer form the cache server in no,
the process continues to step S17, wherein the intercep-
tor determines not to intercept the request from the
user, which is instead passed on to the designated
content provider via the Internet. The process then
continues to step 518, in which the interceptor monitors
the answer sent from the content provider to the end
user. When such information passes the interceptor on its
way to the end user, the interceptor, in step 519, stores
the previously unattainable information in the cache
server 140 and distributes the information to all other
cache servers within the area 200 in Fig 1. Hence, the
received new, previously not stored information will then
be stored in essentially all cache servers within the
area and thus be easily accessed by all users within the
area. Note that every time information is received from
the cache server of from the net, a copy is preferably
temporarily stored in the quick access cache server 112.
Another exemplifying embodiment of an Internet
caching system according to the present invention, simi-
lar to the one shown in Fig 1, will now be described with
reference to Fig 7. In Fig 7, the system differs from the
one shown in Fig 1 in the answer from the content
provider 130e, in this case being located in the region
Eoutside the system area 200, relating to an information
request originally sent by an end user 120a, will not be
transmitted back to the user via the Internet 100.
Instead, the answer is transmitted directly via the
satellite 160 to the servers 140a, 140b, 140c, and 190e,
thereby automatically updating all cache servers 140a.
The cache server 140a and the service provider 110a then
CA 02297658 2000-O1-21
WO 99/05584 PCTlSE98/01316
17
handles the delivery of said information to the end user
120a.
Yet another exemplifying embodiment of an Internet
caching system according to the present invention,
similar to the one shown in Fig 1, will now be described
with reference to Fig 8. In Fig 8, the cache servers
within the area 200, each serving a respective region A,
B, C; and D, is connected to via data channels to a
central control unit 190. The central control unit 190
controls the decisions as to weather or not information
shall be cached in the respective cache servers and as to
how it is to be distributed to other caches within the
area 200. Tn this case, information may be distributed
between the different caches either using the satellite
link or using the data channel connection to and via the
central control unit 190.
Furthermore, with reference primarily to Fig 1, 7
and 8 above, it is understood that decisions regarding
how to handle the caching of information relating to
"local" requests, i.e. for example an information request
from an end user 120a relating to a content provider 130a
located within the same region A, which has been discus-
sed above in the summary of the invention, the different
schemes of operation, including the needed decision-
making, may be implemented either at each of the cache
servers 140, at each of the service providers 110, or at
a central control unit 190, for example.
It is understood that the construction and function
of the elements described with reference to the drawings
will become apparent for those skilled in the art.
Even though the invention has been described with
reference to specific exemplifying embodiments, many
different alterations, modifications and the like will
become apparent for those skilled in the art. The des-
cribed embodiments are therefore not intended to limit
the scope of the invention, as defined by the appended
claims.
